A network device implements a foldable ingress buffer for buffering data units as they are being received. The buffer is organized into a grid of memory banks, having different columns and rows. A Transport Data Unit (“TDU”) is stored interleaved across entries in multiple banks. As each portion of a TDU is received, the portion is written to a different bank of the buffer. In each column of the buffer, a full-sized TDU has portions in a number of rows equal to the number of folds in the buffer. The sum of the bank widths for each row thus needs be no larger than half the maximum TDU size, which further means that the number of columns in the grid of banks may be reduced by at least half compared to non-folded approaches, with little increase in the number of rows, if any, depending on blocking and reading requirements.
Legal claims defining the scope of protection, as filed with the USPTO.
2. The network apparatus of claim 1, wherein each of the memory banks has a width equal to the maximum portion size, and each of the data unit portions is no greater in size than the maximum portion size.
4. The network apparatus of claim 1, wherein each row in the grid is accessible by only one writer or the reader in a given memory access cycle, wherein the number of rows in the buffer is equal to the number of writers, plus the number of folds, plus one.
6. The network apparatus of claim 1, wherein the one or more downstream components include one or more ingress packet processors.
8. The network apparatus of claim 1, wherein each of the memory banks is a different single-ported Static Random-Access Memories (“SRAM”).
9. The network apparatus of claim 1, wherein at least one input of the inputs comprises a multiplexor configured to switch between two or more ports each clock cycle, the input configured to convey data unit portions from alternating ports of the two or more ports in alternating clock cycles.
10. The network apparatus of claim 1, wherein at least one input of the inputs is configured to, in response to receiving both an ending part of a first PDU and a starting part of a second PDU in a single input clock cycle, delay sending the starting part of the second PDU to the writer coupled to the input, the input configured to subsequently convey multiple data unit portions of the second PDU in a single clock cycle, the writer configured to write the multiple data unit portions to different banks in the same row of the grid in a single memory access cycle.
15. The method of claim 11, further comprising, at least one input of the inputs, switching between two or more ports each clock cycle, the input conveying data unit portions from alternating ports of the two or more ports in alternating clock cycles.
16. The method of claim 11, further comprising, in response to an input receiving both an ending part of a first PDU and a starting part of a second PDU in a single input clock cycle, delaying sending the starting part of the second PDU to the writer coupled to the input, the input subsequently conveying multiple data unit portions of the second PDU in a single clock cycle, the writer writing the multiple data unit portions to different banks in the same row of the grid in a single memory access cycle.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
July 20, 2020
January 30, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.